Compare commits

..

76 Commits

Author SHA1 Message Date
ffeaec7fe9 3.0.12 2022-09-16 08:27:25 +02:00
a31e196e5b fix(core): update 2022-09-16 08:27:25 +02:00
57851d90a5 3.0.11 2022-09-15 20:52:22 +02:00
1e4b16b734 fix(core): update 2022-09-15 20:52:22 +02:00
876466b18d 3.0.10 2022-09-13 18:39:38 +02:00
82b1c187ee fix(core): update 2022-09-13 18:39:38 +02:00
4f93d258b8 3.0.9 2022-09-13 18:26:15 +02:00
60993fc005 fix(core): update 2022-09-13 18:26:14 +02:00
db4965c8f5 3.0.8 2022-09-05 01:02:05 +02:00
36a964d931 fix(core): update 2022-09-05 01:02:05 +02:00
6d90f4c9b4 3.0.7 2022-06-29 00:24:44 +02:00
839eafd73f fix(core): update 2022-06-29 00:24:43 +02:00
f4141bf201 3.0.6 2022-06-15 22:16:44 +02:00
5a46028053 fix(core): update 2022-06-15 22:16:44 +02:00
18d26647e1 3.0.5 2022-06-15 22:14:55 +02:00
0b5ec86780 fix(core): update 2022-06-15 22:14:55 +02:00
5e15729045 3.0.4 2022-06-12 19:26:49 +02:00
c98a5f1ac3 fix(core): update 2022-06-12 19:26:49 +02:00
7f02146a0e 3.0.3 2022-06-12 19:26:34 +02:00
5dab8c7351 fix(core): update 2022-06-12 19:26:34 +02:00
505e0e9a30 3.0.2 2022-03-25 01:14:12 +01:00
9e3510955a fix(core): update 2022-03-25 01:14:12 +01:00
c006d4fbc7 3.0.1 2022-03-24 23:32:28 +01:00
5e02c44647 fix(core): update 2022-03-24 23:32:27 +01:00
47f7cb18b5 3.0.0 2022-03-24 14:34:46 +01:00
3faf065c66 2.0.19 2022-03-24 14:32:50 +01:00
dfffd03790 fix(core): update 2022-03-24 14:32:49 +01:00
796aa905d2 2.0.18 2022-01-06 13:23:05 +01:00
253fb95143 fix(core): update 2022-01-06 13:23:04 +01:00
f3ea075b72 2.0.17 2022-01-06 13:10:12 +01:00
af725a7f78 fix(core): update 2022-01-06 13:10:12 +01:00
016e0db797 2.0.16 2022-01-05 23:55:38 +01:00
4cf8b2e1f8 fix(core): update 2022-01-05 23:55:37 +01:00
67b0aa9d47 2.0.15 2022-01-05 17:20:28 +01:00
567c6eafea fix(core): update 2022-01-05 17:20:28 +01:00
ff890fb2af 2.0.14 2022-01-05 16:32:48 +01:00
a512fd64b5 fix(core): update 2022-01-05 16:32:47 +01:00
377318a62a 2.0.13 2022-01-05 14:19:40 +01:00
671c871304 fix(core): update 2022-01-05 14:19:39 +01:00
e0cc6b5655 2.0.12 2022-01-05 14:17:43 +01:00
e74b44b49c fix(core): update 2022-01-05 14:17:43 +01:00
d6f0d88d4a 2.0.11 2021-10-14 16:04:58 +02:00
9674e5b8dc fix(core): update 2021-10-14 16:04:58 +02:00
ea4cf777a9 2.0.10 2021-10-14 10:59:45 +02:00
8a308fa9e3 fix(core): update 2021-10-14 10:59:45 +02:00
56fa53b701 2.0.9 2021-04-29 15:19:15 +00:00
2a4ddd4e41 fix(core): update 2021-04-29 15:19:14 +00:00
3d8a63fddd 2.0.8 2021-04-29 09:32:13 +00:00
9d9b1d0399 fix(core): update 2021-04-29 09:32:12 +00:00
739d60d410 2.0.7 2021-04-14 11:35:06 +00:00
9003034d0d fix(core): update 2021-04-14 11:35:05 +00:00
5002513d21 2.0.6 2021-04-14 11:34:34 +00:00
12ede2be02 fix(core): update 2021-04-14 11:34:33 +00:00
163ec2bd50 2.0.5 2021-03-06 15:33:08 +00:00
7744839613 fix(core): update 2021-03-06 15:33:07 +00:00
df803d90cb 2.0.4 2021-03-05 15:38:12 +00:00
5fc22585bf fix(core): update 2021-03-05 15:38:11 +00:00
b8387458db 2.0.3 2021-03-05 15:16:23 +00:00
04aed2556b fix(core): update 2021-03-05 15:16:22 +00:00
4cde3a7fe0 2.0.2 2019-11-19 16:31:34 +00:00
466d6d47ba fix(ci): further speed up snyk step 2019-11-19 16:31:33 +00:00
fa99b7f068 2.0.1 2019-11-19 16:28:14 +00:00
3e865b6c92 fix(ci): update snyk step for faster testing 2019-11-19 16:28:13 +00:00
94dd576d3b 2.0.0 2019-11-19 15:53:14 +00:00
9c19b4a3e4 BREAKING CHANGE(API): now clearly naming PDF Results what they are. 2019-11-19 15:53:14 +00:00
1155656c2b 1.0.29 2019-11-16 00:50:12 +01:00
5cf23c1134 fix(core): update 2019-11-16 00:50:11 +01:00
67e9fdd10a 1.0.28 2019-11-15 20:00:25 +01:00
b0e388c50d 1.0.27 2019-11-15 19:59:57 +01:00
f37c9d8375 fix(core): update 2019-11-15 19:59:57 +01:00
18c8ef9606 1.0.26 2019-11-12 15:45:26 +01:00
6766a3d0dc fix(core): update 2019-11-12 15:45:26 +01:00
d16f447048 1.0.25 2019-11-12 15:41:58 +01:00
37c752c2ae fix(core): update 2019-11-12 15:41:58 +01:00
b088d26e6f 1.0.24 2019-11-12 14:41:04 +01:00
425f11a334 fix(core): update 2019-11-12 14:41:03 +01:00
21 changed files with 12540 additions and 1706 deletions

4
.gitignore vendored
View File

@ -15,8 +15,6 @@ node_modules/
# builds
dist/
dist_web/
dist_serve/
dist_ts_web/
dist_*/
# custom

View File

@ -19,22 +19,35 @@ mirror:
stage: security
script:
- npmci git mirror
only:
- tags
tags:
- lossless
- docker
- notpriv
snyk:
auditProductionDependencies:
image: registry.gitlab.com/hosttoday/ht-docker-node:npmci
stage: security
script:
- npmci npm prepare
- npmci command npm install -g snyk
- npmci command npm install --ignore-scripts
- npmci command snyk test
- npmci command npm install --production --ignore-scripts
- npmci command npm config set registry https://registry.npmjs.org
- npmci command npm audit --audit-level=high --only=prod --production
tags:
- lossless
- docker
- notpriv
auditDevDependencies:
image: registry.gitlab.com/hosttoday/ht-docker-node:npmci
stage: security
script:
- npmci npm prepare
- npmci command npm install --ignore-scripts
- npmci command npm config set registry https://registry.npmjs.org
- npmci command npm audit --audit-level=high --only=dev
tags:
- docker
allow_failure: true
# ====================
# test stage
@ -49,9 +62,7 @@ testStable:
- npmci npm test
coverage: /\d+.?\d+?\%\s*coverage/
tags:
- lossless
- docker
- priv
testBuild:
stage: test
@ -62,9 +73,7 @@ testBuild:
- npmci command npm run build
coverage: /\d+.?\d+?\%\s*coverage/
tags:
- lossless
- docker
- notpriv
release:
stage: release
@ -84,6 +93,8 @@ release:
codequality:
stage: metadata
allow_failure: true
only:
- tags
script:
- npmci command npm install -g tslint typescript
- npmci npm prepare

9
.snyk
View File

@ -1,9 +0,0 @@
# Snyk (https://snyk.io) policy file, patches or ignores known vulnerabilities.
version: v1.13.5
# ignores vulnerabilities until expiry date; change duration by modifying expiry date
ignore:
SNYK-JS-LODASH-450202:
- '@pushrocks/smartnetwork > portscanner > async > lodash':
reason: None given
expires: '2019-09-13T09:39:26.276Z'
patch: {}

29
.vscode/launch.json vendored Normal file
View File

@ -0,0 +1,29 @@
{
"version": "0.2.0",
"configurations": [
{
"name": "current file",
"type": "node",
"request": "launch",
"args": [
"${relativeFile}"
],
"runtimeArgs": ["-r", "@gitzone/tsrun"],
"cwd": "${workspaceRoot}",
"protocol": "inspector",
"internalConsoleOptions": "openOnSessionStart"
},
{
"name": "test.ts",
"type": "node",
"request": "launch",
"args": [
"test/test.ts"
],
"runtimeArgs": ["-r", "@gitzone/tsrun"],
"cwd": "${workspaceRoot}",
"protocol": "inspector",
"internalConsoleOptions": "openOnSessionStart"
}
]
}

26
.vscode/settings.json vendored Normal file
View File

@ -0,0 +1,26 @@
{
"json.schemas": [
{
"fileMatch": ["/npmextra.json"],
"schema": {
"type": "object",
"properties": {
"npmci": {
"type": "object",
"description": "settings for npmci"
},
"gitzone": {
"type": "object",
"description": "settings for gitzone",
"properties": {
"projectType": {
"type": "string",
"enum": ["website", "element", "service", "npm", "wcc"]
}
}
}
}
}
}
]
}

View File

View File

@ -4,6 +4,7 @@
"npmAccessLevel": "public"
},
"gitzone": {
"projectType": "npm",
"module": {
"githost": "gitlab.com",
"gitscope": "pushrocks",

13788
package-lock.json generated

File diff suppressed because it is too large Load Diff

View File

@ -1,42 +1,52 @@
{
"name": "@pushrocks/smartpdf",
"version": "1.0.23",
"version": "3.0.12",
"private": false,
"description": "create pdfs on the fly",
"main": "dist/index.js",
"typings": "dist/index.d.ts",
"main": "dist_ts/index.js",
"typings": "dist_ts/index.d.ts",
"type": "module",
"author": "Lossless GmbH",
"license": "MIT",
"scripts": {
"test": "(tstest test/)",
"format": "(gitzone format)",
"build": "(tsbuild)"
"test": "(tstest test/ --web)",
"build": "(tsbuild --web --allowimplicitany)"
},
"devDependencies": {
"@gitzone/tsbuild": "^2.1.17",
"@gitzone/tsrun": "^1.2.8",
"@gitzone/tstest": "^1.0.28",
"@pushrocks/tapbundle": "^3.0.13",
"@types/node": "^12.12.7"
"@gitzone/tsbuild": "^2.1.65",
"@gitzone/tsdoc": "^1.1.12",
"@gitzone/tsrun": "^1.2.35",
"@gitzone/tstest": "^1.0.73",
"@pushrocks/tapbundle": "^5.0.4",
"@types/node": "^18.7.18"
},
"dependencies": {
"@pushrocks/smartfile": "^7.0.6",
"@pushrocks/smartnetwork": "^1.1.14",
"@pushrocks/smartpromise": "^3.0.6",
"@pushrocks/smartpuppeteer": "^1.0.5",
"@pushrocks/smartunique": "^3.0.1",
"@types/express": "^4.17.2",
"express": "^4.17.1"
"@pushrocks/smartdelay": "^2.0.13",
"@pushrocks/smartfile": "^10.0.5",
"@pushrocks/smartnetwork": "^3.0.0",
"@pushrocks/smartpath": "^5.0.5",
"@pushrocks/smartpromise": "^3.1.7",
"@pushrocks/smartpuppeteer": "^2.0.2",
"@pushrocks/smartunique": "^3.0.3",
"@tsclass/tsclass": "^4.0.21",
"@types/express": "^4.17.14",
"express": "^4.18.1",
"pdf-merger-js": "^3.4.0",
"pdf2json": "^2.0.0"
},
"files": [
"ts/**/*",
"ts_web/**/*",
"dist/**/*",
"dist_web/**/*",
"dist_*/**/*",
"dist_ts/**/*",
"dist_ts_web/**/*",
"assets/**/*",
"cli.js",
"npmextra.json",
"readme.md"
],
"browserslist": [
"last 1 chrome versions"
]
}

View File

@ -8,16 +8,26 @@ Create PDFs fast and smoothly
* [docs (typedoc)](https://pushrocks.gitlab.io/smartpdf/)
## Status for master
[![build status](https://gitlab.com/pushrocks/smartpdf/badges/master/build.svg)](https://gitlab.com/pushrocks/smartpdf/commits/master)
[![coverage report](https://gitlab.com/pushrocks/smartpdf/badges/master/coverage.svg)](https://gitlab.com/pushrocks/smartpdf/commits/master)
[![npm downloads per month](https://img.shields.io/npm/dm/@pushrocks/smartpdf.svg)](https://www.npmjs.com/package/@pushrocks/smartpdf)
[![Known Vulnerabilities](https://snyk.io/test/npm/@pushrocks/smartpdf/badge.svg)](https://snyk.io/test/npm/@pushrocks/smartpdf)
[![TypeScript](https://img.shields.io/badge/TypeScript->=%203.x-blue.svg)](https://nodejs.org/dist/latest-v10.x/docs/api/)
[![node](https://img.shields.io/badge/node->=%2010.x.x-blue.svg)](https://nodejs.org/dist/latest-v10.x/docs/api/)
[![JavaScript Style Guide](https://img.shields.io/badge/code%20style-prettier-ff69b4.svg)](https://prettier.io/)
Status Category | Status Badge
-- | --
GitLab Pipelines | [![pipeline status](https://gitlab.com/pushrocks/smartpdf/badges/master/pipeline.svg)](https://lossless.cloud)
GitLab Pipline Test Coverage | [![coverage report](https://gitlab.com/pushrocks/smartpdf/badges/master/coverage.svg)](https://lossless.cloud)
npm | [![npm downloads per month](https://badgen.net/npm/dy/@pushrocks/smartpdf)](https://lossless.cloud)
Snyk | [![Known Vulnerabilities](https://badgen.net/snyk/pushrocks/smartpdf)](https://lossless.cloud)
TypeScript Support | [![TypeScript](https://badgen.net/badge/TypeScript/>=%203.x/blue?icon=typescript)](https://lossless.cloud)
node Support | [![node](https://img.shields.io/badge/node->=%2010.x.x-blue.svg)](https://nodejs.org/dist/latest-v10.x/docs/api/)
Code Style | [![Code Style](https://badgen.net/badge/style/prettier/purple)](https://lossless.cloud)
PackagePhobia (total standalone install weight) | [![PackagePhobia](https://badgen.net/packagephobia/install/@pushrocks/smartpdf)](https://lossless.cloud)
PackagePhobia (package size on registry) | [![PackagePhobia](https://badgen.net/packagephobia/publish/@pushrocks/smartpdf)](https://lossless.cloud)
BundlePhobia (total size when bundled) | [![BundlePhobia](https://badgen.net/bundlephobia/minzip/@pushrocks/smartpdf)](https://lossless.cloud)
Platform support | [![Supports Windows 10](https://badgen.net/badge/supports%20Windows%2010/yes/green?icon=windows)](https://lossless.cloud) [![Supports Mac OS X](https://badgen.net/badge/supports%20Mac%20OS%20X/yes/green?icon=apple)](https://lossless.cloud)
## Usage
## Contribution
We are always happy for code contributions. If you are not the code contributing type that is ok. Still, maintaining Open Source repositories takes considerable time and thought. If you like the quality of what we do and our modules are useful to you we would appreciate a little monthly contribution: You can [contribute one time](https://lossless.link/contribute-onetime) or [contribute monthly](https://lossless.link/contribute). :)
## Contribution

View File

@ -1,11 +1,11 @@
import { expect, tap } from '@pushrocks/tapbundle';
import * as smartpdf from '../ts/index';
import * as smartpdf from '../ts/index.js';
let testSmartPdf: smartpdf.SmartPdf;
tap.test('should create a valid instance of smartpdf', async () => {
testSmartPdf = new smartpdf.SmartPdf();
expect(testSmartPdf).to.be.instanceof(smartpdf.SmartPdf);
expect(testSmartPdf).toBeInstanceOf(smartpdf.SmartPdf);
});
tap.test('should start the instance', async () => {
@ -13,30 +13,44 @@ tap.test('should start the instance', async () => {
});
tap.test('should create a pdf from html string', async () => {
await testSmartPdf.getPdfForHtmlString('hi');
await testSmartPdf.getA4PdfResultForHtmlString('hi');
});
tap.test('should create a pdf from html string', async () => {
await testSmartPdf.getA4PdfResultForHtmlString('hi');
});
tap.test('should create a pdf from website as A4', async () => {
await testSmartPdf.getPdfForWebsite('https://maintainedby.lossless.com');
await testSmartPdf.getPdfResultForWebsite('https://www.wikipedia.org');
});
tap.test('should create a pdf from website as single page PDF', async () => {
await testSmartPdf.getFullWebsiteAsSinglePdf('https://maintainedby.lossless.com');
await testSmartPdf.getFullWebsiteAsSinglePdf('https://www.wikipedia.org');
});
tap.test('should create a valid PDFResult', async () => {
const pdfResult = await testSmartPdf.getFullWebsiteAsSinglePdf(
'https://maintainedby.lossless.com'
);
expect(pdfResult.buffer).to.be.instanceOf(Buffer);
const fs = await import('fs');
const writePDfToDisk = async (urlArg: string, fileName: string) => {
const pdfResult = await testSmartPdf.getFullWebsiteAsSinglePdf(urlArg);
expect(pdfResult.buffer).toBeInstanceOf(Buffer);
const fs = await import('fs');
if (!fs.existsSync('.nogit/')) {
fs.mkdirSync('.nogit/');
}
fs.writeFileSync('.nogit/sample.pdf', pdfResult.buffer);
if (!fs.existsSync('.nogit/')) {
fs.mkdirSync('.nogit/');
}
fs.writeFileSync(`.nogit/${fileName}`, pdfResult.buffer as Buffer);
};
await writePDfToDisk('https://maintainedby.lossless.com/', '1.pdf')
await writePDfToDisk('https://rendertron.lossless.one/render/https://layer.io', '2.pdf')
});
tap.test('should combine pdfs', async () => {
const fs = await import('fs');
const pdf1 = await testSmartPdf.readFileToPdfObject('.nogit/1.pdf');
const pdf2 = await testSmartPdf.readFileToPdfObject('.nogit/2.pdf');
fs.writeFileSync(`.nogit/combined.pdf`, (await testSmartPdf.mergePdfs([pdf1, pdf2])).buffer as Buffer);
})
tap.test('should be able to close properly', async () => {
await testSmartPdf.stop();
});

8
ts/00_commitinfo_data.ts Normal file
View File

@ -0,0 +1,8 @@
/**
* autocreated commitinfo by @pushrocks/commitinfo
*/
export const commitinfo = {
name: '@pushrocks/smartpdf',
version: '3.0.12',
description: 'create pdfs on the fly'
}

View File

@ -6,4 +6,4 @@ declare global {
}
// normal
export * from './smartpdf.classes.smartpdf';
export * from './smartpdf.classes.smartpdf.js';

View File

@ -1 +0,0 @@
export * from './interface.pdfresult';

View File

@ -1,5 +0,0 @@
export interface IPdfResult {
name: string;
id: string;
buffer: Buffer;
}

View File

@ -1,8 +1,15 @@
import * as plugins from './smartpdf.plugins';
/*
* pdf candidate is a construct used internally for mapping html
* to pdf buffers delivered by puppeteer
*/
import * as plugins from './smartpdf.plugins.js';
export class PdfCandidate {
pdfId = plugins.smartunique.shortId();
doneDeferred = plugins.smartpromise.defer();
public htmlString: string;
public pdfId = plugins.smartunique.shortId();
public doneDeferred = plugins.smartpromise.defer();
constructor(public htmlString) {}
constructor(htmlStringArg: string) {
this.htmlString = htmlStringArg;
}
}

View File

@ -1,13 +1,19 @@
import * as plugins from './smartpdf.plugins';
import * as paths from './smartpdf.paths';
import * as plugins from './smartpdf.plugins.js';
import * as paths from './smartpdf.paths.js';
import { Server } from 'http';
import { PdfCandidate } from './smartpdf.classes.pdfcandidate';
import { PdfCandidate } from './smartpdf.classes.pdfcandidate.js';
import { IPdf } from '@tsclass/tsclass/dist_ts/business/pdf.js';
declare const document;
import * as interfaces from './interfaces';
declare const document: any;
export class SmartPdf {
// STATIC
public static async create() {
const smartpdfInstance = new SmartPdf();
return smartpdfInstance;
}
// INSTANCE
htmlServerInstance: Server;
serverPort: number;
headlessBrowser: plugins.smartpuppeteer.puppeteer.Browser;
@ -19,14 +25,17 @@ export class SmartPdf {
this._readyDeferred = new plugins.smartpromise.Deferred();
}
async start(headlessBrowserArg?) {
async start(headlessBrowserArg?: plugins.smartpuppeteer.puppeteer.Browser) {
const done = plugins.smartpromise.defer();
// lets set the external browser in case one is provided
this.headlessBrowser = headlessBrowserArg;
// setup puppeteer
if (this.headlessBrowser) {
this.externalBrowserBool = true;
} else {
this.headlessBrowser = await plugins.smartpuppeteer.getEnvAwareBrowserInstance();
this.headlessBrowser = await plugins.smartpuppeteer.getEnvAwareBrowserInstance({
forceNoSandbox: true,
});
}
// setup server
@ -37,11 +46,13 @@ export class SmartPdf {
});
this.htmlServerInstance = plugins.http.createServer(app);
const smartnetworkInstance = new plugins.smartnetwork.SmartNetwork();
const portAvailable = smartnetworkInstance.isLocalPortAvailable(3210);
const portAvailable = smartnetworkInstance.isLocalPortUnused(3210);
this.htmlServerInstance.listen(3210, 'localhost');
this.htmlServerInstance.on('listening', () => {
this._readyDeferred.resolve();
done.resolve();
});
await done.promise;
}
// stop
@ -61,14 +72,19 @@ export class SmartPdf {
/**
* returns a pdf for a given html string;
*/
async getPdfForHtmlString(htmlStringArg: string): Promise<interfaces.IPdfResult> {
async getA4PdfResultForHtmlString(htmlStringArg: string): Promise<plugins.tsclass.business.IPdf> {
await this._readyDeferred.promise;
const pdfCandidate = new PdfCandidate(htmlStringArg);
this._candidates[pdfCandidate.pdfId] = pdfCandidate;
const page = await this.headlessBrowser.newPage();
const response = await page.goto(`http://localhost:3210/${pdfCandidate.pdfId}`, {
waitUntil: 'networkidle2'
await page.setViewport({
width: 794,
height: 1122,
});
const response = await page.goto(`http://localhost:3210/${pdfCandidate.pdfId}`, {
waitUntil: 'networkidle2',
});
// await plugins.smartdelay.delayFor(1000);
const headers = response.headers();
if (headers['pdf-id'] !== pdfCandidate.pdfId) {
console.log('Error! Headers do not match. For security reasons no pdf is being emitted!');
@ -78,7 +94,10 @@ export class SmartPdf {
}
const pdfBuffer = await page.pdf({
format: 'A4'
width: 794,
height: 1122,
printBackground: true,
displayHeaderFooter: false,
});
await page.close();
delete this._candidates[pdfCandidate.pdfId];
@ -87,52 +106,124 @@ export class SmartPdf {
return {
id: pdfCandidate.pdfId,
name: `${pdfCandidate.pdfId}.js`,
buffer: pdfBuffer
metadata: {
textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
},
buffer: pdfBuffer,
};
}
async getPdfForWebsite(websiteUrl: string): Promise<interfaces.IPdfResult> {
async getPdfResultForWebsite(websiteUrl: string): Promise<plugins.tsclass.business.IPdf> {
const page = await this.headlessBrowser.newPage();
page.emulateMedia('screen');
const response = await page.goto(websiteUrl, { waitUntil: 'networkidle2' });
const pdfId = plugins.smartunique.shortId();
const pdfBuffer = await page.pdf({
format: 'A4',
printBackground: true,
displayHeaderFooter: false,
preferCSSPageSize: true
await page.setViewport({
width: 1980,
height: 1200,
});
await page.close();
return {
id: pdfId,
name: `${pdfId}.js`,
buffer: pdfBuffer
};
}
async getFullWebsiteAsSinglePdf(websiteUrl: string) {
const page = await this.headlessBrowser.newPage();
page.emulateMedia('screen');
await page.emulateMediaType('screen');
const response = await page.goto(websiteUrl, { waitUntil: 'networkidle2' });
const pdfId = plugins.smartunique.shortId();
const { documentHeight, documentWidth } = await page.evaluate(() => {
return {
documentHeight: document.height,
documentWidth: document.width
documentWidth: document.width,
};
});
const pdfBuffer = await page.pdf({
height: documentWidth,
height: documentHeight,
width: documentWidth,
printBackground: true,
displayHeaderFooter: false,
preferCSSPageSize: true
});
await page.close();
return {
id: pdfId,
name: `${pdfId}.js`,
buffer: pdfBuffer
metadata: {
textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
},
buffer: pdfBuffer,
};
}
async getFullWebsiteAsSinglePdf(websiteUrl: string): Promise<plugins.tsclass.business.IPdf> {
const page = await this.headlessBrowser.newPage();
await page.setViewport({
width: 1920,
height: 1200,
});
page.emulateMediaType('screen');
const response = await page.goto(websiteUrl, { waitUntil: 'networkidle2' });
const pdfId = plugins.smartunique.shortId();
const { documentHeight, documentWidth } = await page.evaluate(() => {
return {
documentHeight: document.body.scrollHeight,
documentWidth: document.body.clientWidth,
};
});
await page.setViewport({
width: 1920,
height: documentHeight,
});
const pdfBuffer = await page.pdf({
height: documentHeight,
width: 1920,
printBackground: true,
displayHeaderFooter: false,
scale: 1,
pageRanges: '1',
});
await page.close();
return {
id: pdfId,
name: `${pdfId}.js`,
metadata: {
textExtraction: await this.extractTextFromPdfBuffer(pdfBuffer),
},
buffer: pdfBuffer,
};
}
public async mergePdfs(pdfArrayArg: plugins.tsclass.business.IPdf[]): Promise<IPdf> {
const merger = new plugins.pdfMerger();
for (const pdf of pdfArrayArg) {
merger.add(Buffer.from(pdf.buffer));
}
const resultBuffer = await merger.saveAsBuffer();
return {
name: 'mergedPdf',
buffer: resultBuffer,
id: null,
metadata: null
};
}
public async readFileToPdfObject(pathArg: string): Promise<plugins.tsclass.business.IPdf> {
const path = plugins.smartpath.transform.makeAbsolute(pathArg);
const parsedPath = plugins.path.parse(path);
const buffer = await plugins.smartfile.fs.toBuffer(path);
return {
name: parsedPath.base,
buffer,
id: null,
metadata: null
}
}
public async extractTextFromPdfBuffer(pdfBufferArg: Buffer): Promise<string> {
const deferred = plugins.smartpromise.defer<string>();
const pdfParser: any = new plugins.pdf2json();
pdfParser.on('pdfParser_dataReady', (pdfData: any) => {
let finalText = '';
for (const page of pdfData.Pages) {
for (const text of page.Texts) {
for (const letter of text.R) {
finalText = finalText + letter.T;
}
}
}
deferred.resolve(finalText);
});
pdfParser.parseBuffer(pdfBufferArg);
return deferred.promise;
}
}

View File

@ -1,3 +1,3 @@
import * as plugins from './smartpdf.plugins';
import * as plugins from './smartpdf.plugins.js';
export const packageDir = plugins.path.join(__dirname, '../');
export const packageDir = plugins.path.join(plugins.smartpath.get.dirnameFromImportMetaUrl(import.meta.url), '../');

View File

@ -6,14 +6,26 @@ export { http, path };
// @pushrocks
import * as smartfile from '@pushrocks/smartfile';
import * as smartdelay from '@pushrocks/smartdelay';
import * as smartpromise from '@pushrocks/smartpromise';
import * as smartpath from '@pushrocks/smartpath';
import * as smartpuppeteer from '@pushrocks/smartpuppeteer';
import * as smartnetwork from '@pushrocks/smartnetwork';
import * as smartunique from '@pushrocks/smartunique';
export { smartfile, smartpromise, smartpuppeteer, smartunique, smartnetwork };
export { smartfile, smartdelay, smartpromise, smartpath, smartpuppeteer, smartunique, smartnetwork };
// tsclass scope
import * as tsclass from '@tsclass/tsclass';
export {
tsclass
}
// thirdparty
import pdfMerger from 'pdf-merger-js';
// @ts-ignore
import pdf2json from 'pdf2json';
import express from 'express';
export { express };
export { pdfMerger, pdf2json, express };

9
tsconfig.json Normal file
View File

@ -0,0 +1,9 @@
{
"compilerOptions": {
"experimentalDecorators": true,
"target": "ES2022",
"module": "ES2022",
"moduleResolution": "nodenext",
"allowSyntheticDefaultImports": true
}
}

View File

@ -1,17 +0,0 @@
{
"extends": ["tslint:latest", "tslint-config-prettier"],
"rules": {
"semicolon": [true, "always"],
"no-console": false,
"ordered-imports": false,
"object-literal-sort-keys": false,
"member-ordering": {
"options":{
"order": [
"static-method"
]
}
}
},
"defaultSeverity": "warning"
}